MVP & Product·4 min read·September 18, 2025

What 'minimum' actually means in MVP

Most MVPs are not minimum and not viable. After enough of them, we have learned that scoping an MVP is the hard part - and that almost every instinct a founder has about it points the wrong way.

We build a lot of MVPs. The single biggest factor in whether one succeeds is not the engineering, the stack, or the timeline. It is the scoping — and scoping an MVP well goes against almost every instinct a founder brings to it. The word "minimum" is right there in the name, and it is still the part everyone gets wrong, including, often, us, until we say it out loud.

What an MVP is for

An MVP exists to answer a question. Usually: will the people we think want this actually use it? That is it. It is an experiment, and the entire value of an experiment is in how fast and how cheaply it gives you a real answer.

Every feature that does not contribute to answering the question is not a feature of the MVP. It is a delay to the answer, paid for in budget and time, with no return until after the question is answered — and possibly never, because the answer might be no.

This sounds obvious written down. It is genuinely hard to hold onto when you are looking at your own product and every part of it feels essential.

Why "minimum" is so hard

A founder's instincts all push toward more, and the instincts are not stupid — they are just aimed at the wrong stage:

  • "It'll look unfinished." To you, who has seen the whole vision. To a first user seeing it cold, a focused product that does one thing well looks better than a broad one that does ten things vaguely.
  • "We'll need it eventually." Probably. Eventually is after the question is answered. Building it now is committing budget to a future that the experiment has not yet justified.
  • "It's just a small addition." Each one is. They are not small together, and "just one more" is the exact mechanism by which a four-week MVP becomes a four-month one.
  • "Competitors have it." Competitors are past the validation stage. You are not yet. Their feature set is the answer to a question you have not asked yet.

Every one of these is reasonable. Every one of them, followed, makes the MVP slower, more expensive, and worse at its actual job.

How we actually scope it

The exercise we run with clients is subtraction, not addition. We start from the whole vision and ask, of each piece, one question: does the experiment fail without this?

Not "is this good", not "will users like this", not "is this part of the vision" — all yes, all irrelevant right now. Only: if this is not here, can the MVP still answer the question it exists to answer? If yes, it is not in the MVP. It is a fast follow, written down, genuinely not forgotten — just not now.

What survives that filter is usually smaller than the founder expected and almost always the right thing to build.

The things you do not get to cut

"Minimum" applies to scope. It does not apply to the foundations, and this is the distinction that keeps an MVP from being a liability:

  • The data model. Schema is the expensive thing to change later. It gets done properly even on a one-week build.
  • The basics of security. Auth, access control, the obvious protections. "It's just an MVP" is not a security posture.
  • A foundation you can grow on. If the MVP works, you are going to build on it. We pick stacks — Supabase, Postgres — where the MVP is a starting point, not something you have to escape.

Cutting scope is right. Cutting corners on the things that are expensive to fix later is just borrowing against your own future, at a bad rate.

What we would tell you

  • The MVP exists to answer one question. Anything not serving that is delaying the answer.
  • Scope by subtraction. Start from everything, remove what the experiment survives without.
  • Your instincts toward "more" are not wrong, they are early. Eventually is a real time. It is just not now.
  • "Minimum" is about scope, never about foundations. Cut features freely. Do not cut the schema or the security.

The hardest part of building an MVP is not writing it. It is the conversation where you agree, out loud, on how little it can be. Have that conversation properly and the build is almost easy. Skip it and no stack, however fast, will save the timeline.

MVPProductStartupsScoping

Working on something similar?

We have probably hit the exact scaling or architectural bottlenecks you are navigating. Book a call and we'll walk through the technical trade-offs.